M3 MacBook Air Teardown Shows Apple Fixed Base Model's Biggest Flaw

[MacRumors]

The repair website iFixit today shared a video teardown of the base model 13-inch MacBook Air with the M3 chip and 256GB of storage, and it shows that this configuration is equipped with two 128GB flash storage chips. This change results in significantly faster SSD speeds compared to the equivalent MacBook Air with the M2 chip, which has a single 256GB storage chip, as the SSD can read and write from the two chips simultaneously.

YouTube channel Max Tech ran Blackmagic's Disk Speed Test tool with a 5GB file size test on both the M2 and M3 models of the 13-inch MacBook Air with 256GB of storage and 8GB of RAM, and they found the SSD in the M3 model achieved up to 33% faster write speeds and up to 82% faster read speeds compared to the SSD in the M2 model.

Apple's decision to switch to a single 256GB chip for the base model M2 MacBook Air was controversial, even though the slower SSD speeds are unlikely to be noticed by the average user working on common day-to-day tasks. Fortunately, the base model M3 MacBook Air's SSD speeds are now roughly equivalent to the base model M1 MacBook Air again, so customers no longer need to be concerned about this potential limitation.

Apple still sells a 13-inch M2 MacBook Air with 256GB of storage for $999, so customers who want maximum SSD performance should avoid that model.

Beyond this SSD-related change, the teardown shows that the M3 MacBook Air models have a virtually identical internal design as the M2 models. The video provides a look at the battery cells with adhesive pull tabs, logic board, trackpad, and more.

Update: iFixit's CEO Kyle Wiens said the base model 15-inch MacBook Air with the M3 chip and 256GB of storage also has two 128GB storage chips.

Article Link: M3 MacBook Air Teardown Shows Apple Fixed Base Model's Biggest Flaw

 

[brofkand]

I know it's been beaten to death in the year or so since the M2 came out. But I question just how much this means in the real world. Obviously for the price of these machines you don't want any compromises and I get that. But realistically...I didn't buy into the SSD speed boogeyman.

 

[Apple Knowledge Navigator]

I'm not sure if 'fix' is the right term here.

The decision to switch to a single module, just like removing HDMI, SD card ports, and MagSafe from notebooks, was unwarranted to begin with.

They've reversed course and we shouldn't be praising this as an upgrade, but a return to the standard.

 

[Heat_Fan89]

I know it's been beaten to death in the year or so since the M2 came out. But I question just how much this means in the real world. Obviously for the price of these machines you don't want any compromises and I get that. But realistically...I didn't buy into the SSD speed boogeyman.

From my experience, not much! I have a 2018 i5 Mini with a 256GB SSD that was supposedly much faster than the base model M2 Mini SSD.

My base M2 Mini, is twice as fast at everything vs my 2018 Mini. It takes close to 30 secs to cold boot the 2018 Mini and it takes the M2 Mini with the supposedly slow SSD around 9 secs to cold boot.

 

[picpicmac]

Apple Fixed Base Model's Biggest Flaw​

AS IF the M2 is actually flawed in regards to SSDs. Just because one does not like or prefer something does not mean it is flawed.

...and it shows that this configuration is equipped with two 128GB flash storage chips. This change results in significantly faster SSD speeds compared to the equivalent MacBook Air with the M2 chip

A speed difference which is irrelevant to almost all users of an MBA. First off, the "speed" of SSDs is mostly poorly understood; using testing apps, many of the gadflies on the internet only pick the largest number, which is usually for transfers of the biggest blocks. This ignores two things:
- typical use is transfer of small blocks;
- sustained transfers are slower, sometimes much more so, than what is often tested, because the NAND can only write so quickly and a great many people misinterpret speeds reported as if those are NAND speeds, and not system speed. SSDs as used include fast buffers that fill up first.

 

[orbital~debris]

In general use, there's really no problem with this version of the M2 MacBook Air.
Those percentages sound large, but it's relative. In practice, the M2 MacBook Air is extremely speedy and there's a high chance the slower performance won't be noticeable to a majority of users.

 

[thingstoponder]

On principle the m2 shouldn’t have been slower than the m1 but this is so overblown in general

 

[vegetassj4]

"Apple Fixed Base Model's Biggest Flaw"


They fixed the bugs in MacOS....woohooo

oh wait, they are talking about hardware....

:tomatoes, downvotes, and cabbages incoming:

 

[antiprotest]

Apple finally stopped the fraud after years of ripping off customers. Let's keep up the intensity of complaints and lawsuits and public denunciation and ridicule whenever Apple does something shady.

To those who kept defending Apple with "You won't even know the difference," that is the very essence of a scam. But now you will also get the benefit of a faster machine and better value for your money.

Stay vigilant. Don't get Tim cooked.

 

[groove-agent]

So are they RAID 0 disk striping across both SSDs? If so, I wonder if it's hardware or software RAID. Hardware RAID would be sweet.

 

[Fuzzball84]

I know it's been beaten to death in the year or so since the M2 came out. But I question just how much this means in the real world. Obviously for the price of these machines you don't want any compromises and I get that. But realistically...I didn't buy into the SSD speed boogeyman.

Its random access thats far more critical for OS performance.

 

[Fuzzball84]

Apple finally stopped the fraud after years of ripping off customers. Let's keep up the intensity of complaints and lawsuits and public denunciation and ridicule whenever Apple does something shady.

To those who kept defending Apple with "You won't even know the difference," that is the very essence of a scam. But now you will also get the benefit of a faster machine and better value for your money.

Stay vigilant. Don't get Tim cooked.

No one could argue this is fraud 😂

If you dislike the company so much, take your custom elswhere. Its as simple as that…

 

[visualseed]

I know it's been beaten to death in the year or so since the M2 came out. But I question just how much this means in the real world. Obviously for the price of these machines you don't want any compromises and I get that. But realistically...I didn't buy into the SSD speed boogeyman.

it severally impacted the abilities of people that were never going to buy the machine in the first place, so much so, that Apple was forced to fix it so they would stop complaining about it and go back to using their maxed-out MacBook Pros to produce content on other narrowly focused subjects critical of Apple.

 

[Fuzzball84]

On principle the m2 shouldn’t have been slower than the m1 but this is so overblown in general

On what principle? It was a better spec laptop on the whole.

 

[bankshot]

Biggest flaw? I would argue the 8 GB of memory at that price point is a bigger flaw. At a much lower price it would be fine, but at the actual premium starting price, customers should expect more than what was standard 10+ years ago.

 

[Fuzzball84]

AS IF the M2 is actually flawed in regards to SSDs. Just because one does not like or prefer something does not mean it is flawed.



A speed difference which is irrelevant to almost all users of an MBA. First off, the "speed" of SSDs is mostly poorly understood; using testing apps, many of the gadflies on the internet only pick the largest number, which is usually for transfers of the biggest blocks. This ignores two things:
- typical use is transfer of small blocks;
- sustained transfers are slower, sometimes much more so, than what is often tested, because the NAND can only write so quickly and a great many people misinterpret speeds reported as if those are NAND speeds, and not system speed. SSDs as used include fast buffers that fill up first.

There is a whole level of caching going on, writing to single level cell areas before folding into multi bit level areas…. Testing ssd speeds requires understanding of the architecture of the OS, interface, controller and nand etc… most consumers don't want to know, and dont care.

Like you said… marketing and others pick up the big numbers, not knowing truly what they mean in the real world.

 

[sw1tcher]

YouTube channel Max Tech ran Blackmagic's Disk Speed Test tool with a 5GB file size test on both the M2 and M3 models of the 13-inch MacBook Air with 256GB of storage and 8GB of RAM, and they found the SSD in the M3 model achieved up to 33% faster write speeds and up to 82% faster read speeds compared to the SSD in the M2 model.

When asked for comment about the speed improvement, Phil Schiller responded "Can't innovate anymore, my ass!"

 

[Love-hate 🍏 relationship]

>v1 Give people standards (low standards at it , your typical 500$ Dell laptop reaches 6GB up and down )

>V2 Remove speed and lower standards (whilst having people defend it for you) and increase the price (I know the M2 ain't about that but still you get the picture)

>V3 Go back to v1 normalcy (whereas components usually get enhanced after 3 years)

>Profit

 

[visualseed]

AS IF the M2 is actually flawed in regards to SSDs. Just because one does not like or prefer something does not mean it is flawed.



A speed difference which is irrelevant to almost all users of an MBA. First off, the "speed" of SSDs is mostly poorly understood; using testing apps, many of the gadflies on the internet only pick the largest number, which is usually for transfers of the biggest blocks. This ignores two things:
- typical use is transfer of small blocks;
- sustained transfers are slower, sometimes much more so, than what is often tested, because the NAND can only write so quickly and a great many people misinterpret speeds reported as if those are NAND speeds, and not system speed. SSDs as used include fast buffers that fill up first.

You don't understand. Millions of Straw-Man users lost their jobs because they couldn't post Geek Bench scores comparable to peers that bought machines with more storage. It was crazy when HR confronted Tommy two cubes over from me and told him he need to clean out his desk and to take his entry-level POS to the dumpster because the money saved on it was not worth the mental anguish it was causing the other employees waiting the additional microseconds it took for him to complete his work while ultimately affecting the company's bottom line.

 

[zorinlynx]

The real crime here is that they don't start the capacity at 512GB (and also memory at 16GB).

Flash memory and DRAM are both cheap enough now that it's about time they bump the base configuration up to 512GB/16GB.

 

[Jumpthesnark]

The 5GB file transfer test is largely irrelevant to many users on a day to day basis. A more realistic test that would apply to more users would be something that tests different parts of the machine and system, such as downloading a bunch of photos from iCloud, or sending 20 documents to a hosting site (anyone else doing their taxes?).

But I get it. This test is a baseline that (though largely pointless for most) is at least a way to compare the same thing across different machines. And at least in this case, faster read/write speeds is better so yay.

Also I'm glad Apple didn't think "faster transfer speeds are a pro feature so MBAs will keep the single chip configuration and MBPs will have the dual chips." That alone would seem to be proof that Apple did none of this because people here and on YouTube complained, and the change is likely due to supply chain and cost reasons, and nothing more. If M3 Mac buyers are the winners in this then that's great. If Apple did switch back because they heard the complaints, that is also great.

Side note: Those two videos linked in the story couldn't be more different. The iFixit video is so well done and brief. The Max Tech video is awful, with so much wasted time. It would really benefit from an edit that would cut its time in half, and having us be able to scrub through the different parts (like the iFixit video does), but I guess he wants his viewers to have a worse experience, so scrubbing to find what you're interested in is more difficult in his one long video approach.

 

[JPack]

Higher bandwidth also means lower latency. It's a win for all users, but particularly those with 8GB RAM who disk swap when low on memory.

This change is a reversal but the original penny pinching decision shouldn't have occurred in the first place. Apple tried to put as big a gap as possible between MacBook Air and MacBook Pro.

 

[Jumpthesnark]

Biggest flaw? I would argue the 8 GB of memory at that price point is a bigger flaw. At a much lower price it would be fine, but at the actual premium starting price, customers should expect more than what was standard 10+ years ago.

Where does this baseless claim come from, and why do people here continually repeat it, when a 5-second Google search can easily show you that the base RAM of the MBA 10 years ago was 4GB?

 

[MikeSmoke]

Okay. The M2 base might not have been 'flawed' and many users would never know the difference. Truth is there are many circumstances where the difference shows up. Caching, reading and writing audio/video files. A lot of us do lighter 'Pro' work on an Air.

 

[nikhsub1]

I know it's been beaten to death in the year or so since the M2 came out. But I question just how much this means in the real world. Obviously for the price of these machines you don't want any compromises and I get that. But realistically...I didn't buy into the SSD speed boogeyman.

It absolutely is a thing if you get the 8GB ram and the 256GB HD and you use a lot of swap...